Cosine: A Cloud-Cost Optimized Self-Designing Key-Value Storage Engine

Summary: Cosine automatically synthesizes cloud-cost/performance-optimized key-value engines from a 10^36-design space spanning LSMs, B-trees, hash tables, and hybrids. Distribution-aware I/O and learned CPU models enable second-scale search and Rust code generation, yielding up to 53× gains over established engines. (summarized by gpt-5.6-luna on Jul 24 2026)

Paper ID: h27d044d063251329
Venue: VLDB
Year: 2022
Pagerank: 7.2695068e-05
Overall Rank: 3,479 | 76.62%
DOI: 10.14778/3485450.3485461

Incoming Non-self Citations Over Time

Authors

1. Subarna Chatterjee (Harvard University)
2. Meena Jagadeesan (Harvard University)
3. Wilson Qin (Harvard University)
4. Stratos Idreos (Harvard University)

BibTeX Citation

@article{chatterjee_vldb22,
        title = {{Cosine: A Cloud-Cost Optimized Self-Designing Key-Value Storage Engine}},
        author = {Chatterjee, Subarna and Jagadeesan, Meena and Qin, Wilson and Idreos, Stratos},
        journal = {PVLDB},
        series = {{VLDB} '22},
        volume = {15},
        number = {1},
        pages = {112--126},
        doi = {10.14778/3485450.3485461},
        url = {https://doi.org/10.14778/3485450.3485461},
        year = {2022}
}

Incoming Citations (Sorted by Pagerank)

Showing 21 of 21 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
3,232	TreeLine: An Update-In-Place Key-Value Store for Modern Storage	2023	VLDB	7.50343e-05
4,750	GRF: A Global Range Filter for LSM-Trees with Shape Encoding	2024	SIGMOD	6.4354422e-05
5,046	Dissecting, Designing, and Optimizing LSM-based Data Stores	2022	SIGMOD	6.3000825e-05
6,124	Learning to Optimize LSM-trees: Towards A Reinforcement Learning based Key-Value Store for Dynamic Workloads	2023	SIGMOD	5.8788211e-05
7,598	SageDB: An Instance-Optimized Data Analytics System	2022	VLDB	5.4871733e-05
7,905	CAMAL: Optimizing LSM-trees via Active Learning	2024	SIGMOD	5.4291824e-05
7,923	NOCAP: Near-Optimal Correlation-Aware Partitioning Joins	2023	SIGMOD	5.4260253e-05
8,221	ArceKV: Towards Workload-driven LSM-compactions for Key-Value Store Under Dynamic Workloads	2026	VLDB	5.3759492e-05
8,329	How to Grow an LSM-tree? Towards Bridging the Gap Between Theory and Practice	2025	SIGMOD	5.3536854e-05
8,620	Limousine: Blending Learned and Classical Indexes to Self-Design Larger-than-Memory Cloud Storage Engines	2024	SIGMOD	5.301557e-05
9,081	MirrorKV: An Efficient Key-Value Store on Hybrid Cloud Storage with Balanced Performance of Compaction and Querying	2023	SIGMOD	5.2283159e-05
9,130	Rethinking The Compaction Policies in LSM-trees	2025	SIGMOD	5.2233025e-05
9,149	Towards Systematic Index Dynamization	2024	VLDB	5.2186338e-05
9,181	Structural Designs Meet Optimality: Exploring Optimized LSM-tree Structures in A Colossal Configuration Space	2024	SIGMOD	5.2118872e-05
9,646	Are Joins over LSM-trees Ready? Take RocksDB as an Example	2025	VLDB	5.1453267e-05
9,706	Database Gyms	2023	CIDR	5.1376763e-05
10,043	The Image Calculator: 10x Faster Image-AI Inference by Replacing JPEG with Self-designing Storage Format	2024	SIGMOD	5.0921006e-05
10,758	Dynamic read & write optimization with TurtleKV	2026	VLDB	4.9793485e-05
11,430	AXE: A Task Decomposition Approach to Learned LSM Tuning	2025	VLDB	4.9793485e-05
11,555	Breathing New Life into An Old Tree: Resolving Logging Dilemma of B+-tree on Modern Computational Storage Drives	2024	VLDB	4.9793485e-05
11,863	Workload-Adaptive Filtering in Storage Engines	2022	SIGMOD	4.9793485e-05

Outgoing Citations (Sorted by Pagerank)

Showing 29 of 29 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
1	Access Path Selection in a Relational Database Management System	1979	SIGMOD	0.0023947656
40	The Case for Learned Index Structures	2018	SIGMOD	0.00046284649
78	Automatic Database Management System Tuning Through Large-scale Machine Learning	2017	SIGMOD	0.00036684414
224	Self-Driving Database Management Systems	2017	CIDR	0.00024013745
236	LinkBench: a Database Benchmark Based on the Facebook Social Graph	2013	SIGMOD	0.00023671522
253	Database Cracking	2007	CIDR	0.00023042111
272	An Overview of Query Optimization in Relational Systems	1998	PODS	0.00022509573
400	Monkey: Optimal Navigable Key-Value Store	2017	SIGMOD	0.00019129175
546	Faster: A Concurrent Key-Value Store with In-Place Updates	2018	SIGMOD	0.00016590738
555	SageDB: A Learned Database System	2019	CIDR	0.00016506678
1,516	Automatically Indexing Millions of Databases in Microsoft Azure SQL Database	2019	SIGMOD	0.00010402594
1,600	SQLGraph: An Efficient Relational-Based Property Graph Store	2015	SIGMOD	0.00010112271
1,605	The Data Calculator*: Data Structure Design and Cost Synthesis from First Principles and Learned Cost Models	2018	SIGMOD	0.00010093796
1,889	Design Continuums and the Path Toward Self-Designing Key-Value Stores that Know and Learn	2019	CIDR	9.4273689e-05
2,664	Rosetta: A Robust Space-Time Optimized Range Filter for Key-Value Stores	2020	SIGMOD	8.1569551e-05
2,770	Constructing and Analyzing the LSM Compaction Design Space	2021	VLDB	8.037607e-05
2,788	Lethe: A Tunable Delete-Aware LSM Engine	2020	SIGMOD	8.0133966e-05
2,966	Choosing A Cloud DBMS: Architectures and Tradeoffs	2019	VLDB	7.8039817e-05
3,028	Autoscaling Tiered Cloud Storage in Anna	2019	VLDB	7.7377953e-05
3,032	iBTune: Individualized Buffer Tuning for Large-scale Cloud Databases	2019	VLDB	7.7351139e-05
3,275	ForkBase: An Efficient Storage Engine for Blockchain and Forkable Applications	2018	VLDB	7.4691437e-05
3,525	Key-Value Storage Engines	2020	SIGMOD	7.2327042e-05
3,599	Access Path Selection in Main-Memory Optimized Data Systems: Should I Scan or Should I Probe?	2017	SIGMOD	7.1773938e-05
3,641	Nova-LSM: A Distributed, Component-based LSM-tree Key-value Store	2021	SIGMOD	7.1439198e-05
4,772	LogKV: Exploiting Key-Value Stores for Event Log Processing	2013	CIDR	6.4229185e-05
5,397	Order-Preserving Key Compression for In-Memory Search Trees	2020	SIGMOD	6.1490013e-05
6,080	From Auto-tuning One Size Fits All to Self-designed and Learned Data-intensive Systems	2019	SIGMOD	5.8924903e-05
6,635	LSM-Trees and B-Trees: The Best of Both Worlds	2019	SIGMOD	5.7263609e-05
8,044	nKV in Action: Accelerating KV-Stores on Native Computational Storage with Near-Data Processing	2020	VLDB	5.4009328e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	816	SlimDB: A Space-Efficient Key-Value Storage Engine For Semi-Sorted Data	2017	VLDB
2	9,081	MirrorKV: An Efficient Key-Value Store on Hybrid Cloud Storage with Balanced Performance of Compaction and Querying	2023	SIGMOD
3	10,658	Making LSM-Tree-based Key-Value Store Practical and Efficient for Multi-Tenant Serverless Cloud Databases	2026	SIGMOD
4	2,834	Cost Models for Big Data Query Processing: Learning, Retrofitting, and Our Findings	2020	SIGMOD
5	12,534	COCCUS: Self-Configured Cost-Based Query Services in the Cloud	2013	SIGMOD
6	10,933	SOS: A High-Performance Distributed Key-Value Store for Large-Scale Online Services	2026	VLDB
7	10,035	From Worst-Case to Average-Case Analysis: Accurate Latency Predictions for Key-Value Storage Engines	2020	SIGMOD
8	7,241	CaaS-LSM: Compaction-as-a-Service for LSM-based Key-Value Stores in Storage Disaggregated Infrastructure	2024	SIGMOD
9	3,525	Key-Value Storage Engines	2020	SIGMOD
10	8,620	Limousine: Blending Learned and Classical Indexes to Self-Design Larger-than-Memory Cloud Storage Engines	2024	SIGMOD